Background
MicroRNAs (miRNAs) are small (mostly 18–21 nucleotides long), non-coding RNAs, highly conserved across evolution and involved in the regulation of messenger RNAs (mRNAs). Intracellular miRNAs cause post-transcriptional repression of multiple mRNAs to which they bind [
1]. The
let-
7 family of miRNAs in humans consists of twelve genes that encode nine mature miRNAs (
let-
7a,
let-
7b,
let-
7c,
let-
7d,
let-
7e,
let-
7f,
let-
7g,
let-
7i and
miR-
98). The expression of mature
let-
7 miRNAs can be regulated at the transcriptional and post-transcriptional levels, where post-transcriptional repression of
let-
7 (at both
pri-
let-
7 and
pre-
let-
7 stages) is mainly mediated by the RNA-binding protein LIN28 [
2]. Two human homologs of the
C. elegans lin28 gene were identified and named
LIN28A and
LIN28B [
3]. Interestingly, LIN28 proteins and their
let-
7 miRNA targets have several reported functions including regulation of developmental timing [
4‐
7], stem cell pluripotency, and differentiation of skeletal muscle [
8,
9].
In humans, reticulocyte levels of
let-
7 miRNAs increase with the fetal-to-adult developmental transition [
10].
LIN28B expression is silenced during the same developmental switch. Transgenic increases of LIN28 proteins in adult erythroblasts, which as a consequence down-regulate the
let-
7 miRNAs, cause the cells to manifest fetal-like features [
11,
12]. Augmented expression of
LIN28A/B also precipitated a rise in fetal hemoglobin (HbF) levels and amelioration of the sickling morphologies of enucleated erythrocytes cultured in vitro from pediatric patients with sickle cell disease (HbSS genotype) [
11,
12]. Earlier efforts aimed toward the reduced expression of
let-
7 by “sponge” targeting of the miRNA family seed region [
11] resulted in mild HbF increases compared with LIN28 over-expression in the same cells [
11]. Therefore, it remained inconclusive whether suppression of the
let-
7 family, or targeting of individual
let-
7 species are sufficient to cause the robust developmentally-specific changes in cellular phenotype that were manifested by LIN28 [
11].
Here we investigate the expression levels of the individual let-7 miRNAs in human blood cells, and further explore the role of let-7 miRNAs upon ontogeny-related gene expression in the erythroid lineage. Expression of individual let-7 miRNA family members was quantitated in human peripheral blood cell populations, allowing a more focused strategy for reducing let-7 levels in the adult erythroblasts. Finally, lentiviruses designed specifically for let-7a and let-7b miRNA targeting were transduced in erythroblasts and explored for their regulation of HbF and other developmentally-regulated genes.
Methods
Ethics statement
Written informed consent was obtained from all research subjects prior to participation in this study. Approval for the research protocol and consent documents using primary erythroblasts and peripheral blood samples was granted by the Intramural National Institute of Diabetes and Digestive and Kidney Diseases Institutional Review Board.
miRBase
Mature
let-
7 miRNAs sequences were obtained from the miRBase database release 21 (
http://mirbase.org). Details of the miRBase database have been previously described [
13‐
17].
Peripheral blood samples
Peripheral blood cells were isolated using Ficoll-Paque Premium (GE Healthcare, Pittsburgh, PA) following manufacturer’s instructions. Fresh post-ficoll peripheral blood cells were used for cell sorting based on forward and side scatter using the BD FACSAria I flow cytometer (BD Biosciences, San Jose, CA). Lymphocytes and monocytes were sorted from the post-ficoll interface. Neutrophils were obtained from the post-ficoll packed red cells, after lysis with ACK lysing buffer following manufacturer’s protocol (Life Technologies, Grand Island, NY) and sorted based on forward and side scatter. Reticulocytes were obtained after filtration through a Purecell Neonatal High Efficiency Leukocyte Reduction Filter (PALL, Port Washington, NY) of the post-ficoll packed red cells. RNA from lymphocytes, monocytes and neutrophils was extracted using miRNeasy mini kit with QIAzol (Qiagen, Germantown, MD) and RNA from reticulocytes was extracted using Trizol LS following manufacturer’s instructions (Thermo Fisher Scientific, Grand Island, NY).
Cell culture
Cryopreserved healthy adult human CD34(+) cells were cultured ex vivo in a 3-week serum-free system consisting of three phases (phase I: days 0–7, phase II: days 7–14 and phase III: days 14–21) as previously described [
11,
18].
Recombinant viral transduction
Lentiviral particles with tough decoy (TuD) design [
19], constructed to inhibit human
let-
7a or
let-
7b miRNAs (catalog numbers: HLTUD0001 and HLTUD0007, respectively) and negative vector control (HLTUD001C) were purchased from Sigma Aldrich (St. Louis, MO). A lentivirus shRNA vector to knockdown
BCL11A (clone TRCN0000033449) and the respective lentiviral control (SHC002V) were also acquired from Sigma Aldrich. On culture day 3 of phase I, CD34(+) cells were transduced with the following lentiviral particles: let-7a-TuD, let-7b-TuD, BCL11A knockdown and each respective negative vector control (MOI of 6). After 24 h, puromycin (Sigma Aldrich) was added to the culture. On culture day 7, cells were transferred to phase II medium containing EPO and cultivated at the conditions previously described without puromycin [
18].
Cell counts and cell morphology analyses
Cell counts were performed throughout the culture period in a Z1 Coulter Particle Counter (Beckman Coulter, Indianapolis, IN) following manufacturer’s instructions. Cell morphology was analyzed with the preparation of cytospins followed by Wright–Giemsa staining. Briefly, cytospins were prepared by centrifugation of the cytoslides using the Shandon Cytospin 4 (Thermo Fisher Scientific) at 1000 rpm for 2 min. Cytoslides were stained with Wright–Giemsa (Sigma-Aldrich, St. Louis, MO) for 50 s followed by two 1-min washes in distilled water.
Flow cytometry analyses
Erythroid differentiation was assessed with antibodies directed against CD71 and glycophorin A (Invitrogen, Carlsbad, CA) on culture days 14 and 21 using the BD FACSAria I flow cytometer (BD Biosciences) as previously described [
20]. Enucleation was quantitated by thiazole orange (TO) staining (Sigma) on culture day 21. Fetal hemoglobin distribution was assessed with antibody directed against fetal hemoglobin (Life Technologies) at culture day 21 as previously described [
21].
Quantitative PCR for mRNAs
Total RNA was isolated using miRNeasy mini kit with QIAzol (Qiagen) following manufacturer’s instructions and complementary DNA (cDNA) was synthesized using SuperScript III reverse transcriptase (Thermo Fisher) following manufacturer’s instructions as previously described [
22,
23]. RT-qPCR assays and conditions were performed as previously described [
11,
22‐
25]. Assay-on-Demand Gene Expression Product (Thermo Fisher Scientific/Applied Biosystems) were used as follows:
CA1 (Hs01100176_m1),
GCNT2 (Hs00377334_m1),
BCL11A (Hs00256254_m1),
HMGA2 (Hs00971724_m1),
ZBTB7A (Hs00792219_m1),
KLF1 (Hs00610592_m1),
SOX6 (Hs00264525_m1),
LIN28A (Hs04189307_g1), and
LIN28B (Hs01013729_m1). Absolute quantification for each target mRNA was determined by comparison with a standard curve that was run in parallel with biological samples as previously described [
23]. Reactions were performed in triplicate.
Quantitative PCR analysis for the let-7 family of miRNAs
Complementary DNA and real-time PCR reaction using Taqman microRNA assay (Applied Biosystems, Grand Island, NY) were performed as previously described [
10,
22] for
let-
7a,
let-
7b,
let-
7c,
let-
7d,
let-
7e,
let-
7f,
let-
7
g,
let-
7i and
miR-
98. Absolute quantification for each target miRNA was determined by comparison with a standard curve that was run in parallel with biological samples as previously described [
22]. Standard curves were prepared on the basis of the synthetic targeted mature miRNA oligonucleotide of known concentration (at least five 1:10 serial dilutions) as previously described [
22]. Reactions were performed in triplicate. A representative standard curve and its correspondent amplification plot for each
let-
7 miRNA family member is shown in Additional file
1.
Western blot analyses
Nuclear and cytoplasmic extracts from culture day 14 erythroblasts were prepared using the NE-PER Nuclear and Cytoplasmic Extraction kit (Pierce Biotechnology, Rockford, IL) as previously described [
11]. Western blot protocols and conditions were performed as previously described [
11]. Blots were probed with antibodies against CA1 (Abcam, Cambridge, MA), GCNT2 (Santa Cruz Biotechnology, Dallas, TX), BCL11A (Abcam), HMGA2 (GeneTex, Irvine, CA), ZBTB7A (Abcam), KLF1 (Abcam) and SOX6 (Santa Cruz Biotechnology). Histone H3, Lamin B1 or Beta-Actin (all from Abcam) were used as loading controls.
CD34(+) cells from three independent donors were transduced with
let-
7a tough decoy vector (catalog number: HLTUD0001, Sigma) or negative vector control (catalog number: HLTUD001C, Sigma) overnight and then mixed in MethoCult H4034 Optimum media (Stem Cell Technologies, Vancouver, Canada) supplemented with puromycin for colony formation assay with duplicate wells following manufacturer’s protocol as previously described [
18]. Colonies of erythroid progenitors (BFU-E and CFU-E), granulocyte–macrophage progenitors (CFU-GM, CFU-G and CFU-M) and multipotential granulocyte, erythroid, macrophage, megakaryocyte progenitors (CFU-GEMM) were counted for each donor and condition on culture day 14.
HPLC for adult and fetal hemoglobins
Samples for HPLC analysis were prepared and analyzed as previously described [
23,
26].
Statistical analysis
Replicates are expressed as mean ± SD values and significance was calculated by two-tailed Student’s t-test.
Discussion
In this study, we show that the let-7 family of miRNAs is differentially expressed in purified adult human blood, and that let-7a and let-7b are the predominantly expressed family members in the analyzed peripheral blood cell populations, including reticulocytes. Focused suppression of let-7a and let-7b miRNAs with a miRNA Tough Decoy approach in erythroblasts was sufficient to cause robust changes in several developmentally-specific erythroblast genes including increases in gamma-globin mRNA expression and HbF to reach mean levels around 35–40% of the total hemoglobin produced. As such, these data confirm a functional role for erythroblast let-7 miRNAs in globin gene regulation and suggest that targeted reductions of the predominant let-7s should be further explored for application in patients with sickle cell disease and beta-hemoglobin disorders.
The first description of small-RNAs as regulators of developmental timing events was observed in studies of
C. elegans, when
lin-
4 and, subsequently
let-
7 (from the initial denomination
lethal-
7) [
39‐
41] were identified. The mature
let-
7 miRNAs sequence and its function as a heterochronic regulator are highly conserved across evolution [
42], and
let-
7a is the most well-conserved
let-
7 family member. Mature
let-
7a originated from three different genomic loci (
let-
7a-
1,
let-
7a-
2 and
let-
7a-
3) and mature
let-
7f originated from miRNA precursors of two distinct genomic locations (
let-
7f-
1 and
let-
7f-
2), while all other family members are originated from one precursor miRNA sequence.
Also of interest, we observed that
let-
7a and
let-
7b are the major species detected by RT-qPCR in all peripheral blood cell populations analyzed. As reported previously, array-based analyses of the
let-
7 miRNAs demonstrated similarly high levels of each
let-
7 family member in human adult reticulocytes rather than the predominance of
let-
7a and
let-
7b [
10]. The differences between these microarrays
versus RT-qPCR results may be due to a higher level of cross-reactivity in
let-
7 miRNAs array-based detection [
43].
Here we aimed our gene transduction studies to suppress the two most prevalent members of the
let-
7 miRNA family,
let-
7a and
let-
7b. Importantly, the high similarity among the mature
let-
7 miRNA sequences prevented the exclusive targeting by tough decoy (TuD) lentiviral designs. While TuD inhibitors are able to provide a more focused inhibition of miRNAs than other strategies [
27,
28], similar non-specificity of TuD constructs for targeted species in the same miRNA family was previously described [
19]. These results support the notion that targeted
let-
7 inhibition is a robust approach towards the manipulation of HbF levels in adult erythroblasts. Genomic targeting of individual
let-
7 species, perhaps with short palindromic repeat technologies, may be useful for determination of individual
let-
7 family member effects upon HbF.
Interestingly, the magnitude of total
let-
7 suppression was proportional to the increase in
gamma-
globin mRNA and HbF in studies reported to date. It is known that LIN28 proteins regulate
let-
7 biogenesis and that
let-
7 miRNAs regulate
LIN28 levels by binding to its 3′ untranslated region in a double negative feedback loop [
44]. The absence of increased
LIN28A or
LIN28B mRNA transcripts after
let-
7 suppression suggests that the
LIN28 genes are transcriptionally silent rather than post-transcriptionally degraded by
let-
7 in the adult cells. Our study also provides the first evidence that the
let-
7 effects extend beyond globin gene regulation to include other markers of the fetal-to-adult switch in the erythroid lineage, namely
CA1 and
GCNT2. Interestingly, reduced levels were observed only at the protein levels of CA1, while GCNT2 levels remained unchanged. The biological significance of this finding will require further investigation.
The erythroid-related genes
BCL11A,
HMGA2,
ZBTB7A,
KLF1 and
SOX6 were investigated upon
let-
7 suppression. The B-cell CLL/lymphoma 11A (
BCL11A) is a zinc-finger transcription factor known to regulate
gamma-
globin and HbF levels in human erythroid cells [
32] as well as to rescue the sickle cell disease phenotype in a murine model through the activation of HbF [
33]. The High Mobility Group AT-hook 2 (
HMGA2) is known as an architectural transcription factor and a target of the
let-
7 miRNAs [
34] that has been recently reported to regulate
gamma-
globin mRNA and moderately increase HbF levels in human adult erythroblasts in vitro [
22]. The Leukemia/Lymphoma-Related Factor (LRF) encoded by the Zinc Finger and BTB Domain Containing 7A (
ZBTB7A) gene is also a zinc-finger transcription factor shown to cause robust increases in the HbF levels in human cultured erythroblasts [
35]. The Kruppel like factor 1 (
KLF1) is an erythroid-specific transcription factor known to regulate the expression of several erythroid genes including
BCL11A [
36]. Finally, the SRY-box 6 (
SOX6) is a transcription factor that contains a conserved DNA-binding domain and was demonstrated to physically interact and co-occupy the human beta-globin cluster with BCL11A and other transcription factors such as GATA1 [
37]. Interestingly, marked modulations were observed only at BCL11A (down-regulation at both the mRNA and protein levels) and HMGA2 (up-regulation at the protein level).
Overall, we interpret our data as demonstrating that reduction of
let-
7, in the absence of other potential LIN28 effects, is a main driver of these developmentally-regulated genes in erythroblasts. While LIN28 effects upon the expression of
BCL11A have been inconsistent in prior studies [
11,
12], this study shows that robust
let-
7 reduction is sufficient to reduce BCL11A as well as to increase HMGA2 for increased
gamma-
globin transcription. Future studies should be aimed toward understanding how this well-conserved miRNA family is able to regulate erythroid gene activity associated with the fetal-to-adult transition in humans. Since
let-
7 has a more generic role in timing worm development, such studies may ultimately demonstrate how the
let-
7 developmental clock circuit functionally evolved in human tissues.
Authors’ contributions
JFV designed, performed, and analyzed experiments, and wrote the paper. CB, YTL, JMA and MK performed and analyzed experiments. AR conducted clinical research. JLM conceived, assisted, and directed the research, and wrote the paper. All authors read and approved the final manuscript.